/usr/share/pyshared/ase/io/xsf.py is in python-ase 3.6.0.2515-1.1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 | import numpy as np
from ase.atoms import Atoms
from ase.units import Hartree
from ase.parallel import paropen
from ase.calculators.singlepoint import SinglePointCalculator
def write_xsf(fileobj, images, data=None):
if isinstance(fileobj, str):
fileobj = paropen(fileobj, 'w')
if not isinstance(images, (list, tuple)):
images = [images]
fileobj.write('ANIMSTEPS %d\n' % len(images))
numbers = images[0].get_atomic_numbers()
pbc = images[0].get_pbc()
if pbc[2]:
fileobj.write('CRYSTAL\n')
elif pbc[1]:
fileobj.write('SLAB\n')
elif pbc[0]:
fileobj.write('POLYMER\n')
else:
fileobj.write('MOLECULE\n')
for n, atoms in enumerate(images):
if pbc.any():
fileobj.write('PRIMVEC %d\n' % (n + 1))
cell = atoms.get_cell()
for i in range(3):
fileobj.write(' %.14f %.14f %.14f\n' % tuple(cell[i]))
fileobj.write('PRIMCOORD %d\n' % (n + 1))
# Get the forces if it's not too expensive:
calc = atoms.get_calculator()
if (calc is not None and
(hasattr(calc, 'calculation_required') and
not calc.calculation_required(atoms,
['energy', 'forces', 'stress']))):
forces = atoms.get_forces()
else:
forces = None
pos = atoms.get_positions()
fileobj.write(' %d 1\n' % len(pos))
for a in range(len(pos)):
fileobj.write(' %2d' % numbers[a])
fileobj.write(' %20.14f %20.14f %20.14f' % tuple(pos[a]))
if forces is None:
fileobj.write('\n')
else:
fileobj.write(' %20.14f %20.14f %20.14f\n' % tuple(forces[a]))
if data is None:
return
fileobj.write('BEGIN_BLOCK_DATAGRID_3D\n')
fileobj.write(' data\n')
fileobj.write(' BEGIN_DATAGRID_3Dgrid#1\n')
data = np.asarray(data)
if data.dtype == complex:
data = np.abs(data)
shape = data.shape
fileobj.write(' %d %d %d\n' % shape)
cell = atoms.get_cell()
origin = np.zeros(3)
for i in range(3):
if not pbc[i]:
origin += cell[i] / shape[i]
fileobj.write(' %f %f %f\n' % tuple(origin))
for i in range(3):
fileobj.write(' %f %f %f\n' %
tuple(cell[i] * (shape[i] + 1) / shape[i]))
for x in range(shape[2]):
for y in range(shape[1]):
fileobj.write(' ')
fileobj.write(' '.join(['%f' % d for d in data[x, y]]))
fileobj.write('\n')
fileobj.write('\n')
fileobj.write(' END_DATAGRID_3D\n')
fileobj.write('END_BLOCK_DATAGRID_3D\n')
def read_xsf(fileobj, index=-1, read_data=True):
if isinstance(fileobj, str):
fileobj = open(fileobj)
readline = fileobj.readline
while True:
line = readline()
if line[0] != '#':
line = line.strip()
break
if 'ANIMSTEPS' in line:
nimages = int(line.split()[1])
line = readline().strip()
else:
nimages = 1
if 'CRYSTAL' in line:
pbc = True
elif 'SLAB' in line:
pbc = (True, True, False)
elif 'POLYMER' in line:
pbc = (True, False, False)
else:
pbc = False
images = []
for n in range(nimages):
cell = None
if pbc:
line = readline().strip()
assert 'PRIMVEC' in line
cell = []
for i in range(3):
cell.append([float(x) for x in readline().split()])
line = readline().strip()
assert 'PRIMCOORD' in line
natoms = int(readline().split()[0])
numbers = []
positions = []
for a in range(natoms):
line = readline().split()
numbers.append(int(line[0]))
positions.append([float(x) for x in line[1:]])
positions = np.array(positions)
if len(positions[0]) == 3:
forces = None
else:
positions = positions[:, :3]
forces = positions[:, 3:] * Hartree
image = Atoms(numbers, positions, cell=cell, pbc=pbc)
if forces is not None:
image.set_calculator(SinglePointCalculator(None, forces, None,
None, image))
images.append(image)
if read_data:
line = readline()
assert 'BEGIN_BLOCK_DATAGRID_3D' in line
line = readline()
assert 'BEGIN_DATAGRID_3D' in line
shape = [int(x) for x in readline().split()]
start = [float(x) for x in readline().split()]
for i in range(3):
readline()
n_data = shape[0]*shape[1]*shape[2]
data = np.array([float(readline())
for s in range(n_data)]).reshape(shape[::-1])
data = np.swapaxes(data, 0, 2)
return data, images[index]
return images[index]
|